All the pretrained models are available at https://huggingface.co/collections/Keylab/morganbert-681eb0b2479a016a7b7becce
The corpus of compounds for pretraining and validation: https://huggingface.co/datasets/Keylab/MorganBERT_corpus_compound
Downstream task datasets: https://netknowledge.github.io/ADMET/